REVIEW 5 major objections 4 minor 35 references
Path Integral Methods for Synthesizing and Preventing Stealthy Attacks in Nonlinear Cyber-Physical Systems
T0 review · 5 major / 4 minor · reviewed 2026-08-16 · deepseek-v4-flash
Pith's one-line read Path sampling reveals both the stealthy attack and the control that defeats it
desk verdict The attack-synthesis reduction is genuinely clean, but the sampling formulas require an invertibility the paper's own examples lack, so the numerical claims are not reproducible as written. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The load-bearing device is the logarithmic (Cole-Hopf) transformation of the value function, $V_t = \pm\gamma\log\Psi_t$, combined with the Feynman-Kac lemma. Applying this transformation to the HJB equation of the attack problem, or to the HJI equation of the game, cancels the quadratic gradient term exactly when the problem has the right special structure, leaving a linear backward Chapman-Kolmogorov PDE whose solution is an expectation over simulated paths. For attack synthesis the cancellation needs no extra structural condition; for mitigation it requires Assumption 2 or 3, which makes $h_th_t^\top$ proportional to $g_t R_t^{-1} g_t^\top$ (with a $\frac{1}{\lambda}h_th_t^\top$ correction in Assumption 3). The resulting formulas, such as equation (22) for the attack and equations (43)-(44) for the controller and attacker, are weighted averages of noise increments $h_t dw_t$ over trajectories weighted by $\exp(\pm\frac{1}{\gamma}\int \ell_s\,ds)$, computed by Monte Carlo simulation.
What would settle it
Take a two-state system with $g_t = [1;0]$, $h_t = [0;1]$, and $R_t = I$. Then $h_th_t^\top = \mathrm{diag}(0,1)$ while $g_t R_t^{-1} g_t^\top = \mathrm{diag}(1,0)$, so no $\alpha > 0$ satisfies Assumption 3. Solve the HJI equation (47) directly for this system and compare its value to the Monte Carlo estimate (42); a mismatch would confirm that the matching condition is necessary for the mitigation formula, while the attack-synthesis formula (22) should still reproduce the true HJB value.
Extended reading notes
Core claim
The central claim is that both sides of the stealthy attack problem reduce to path integrals over simulated noise paths. For attack synthesis (Problem 1), Theorem 1 shows the KL-constrained attack objective is equivalent to a quadratic-cost stochastic optimal control problem whose optimal bias is $\theta_t^* = (1/\lambda) h_t^\top \partial_x V_t(x_t)$; Theorem 2 evaluates the value function via the Feynman-Kac representation as $V_t = \lambda \log \mathbb{E}_Q[\exp(\frac{1}{\lambda}\int_t^T c_s\,ds)]$ and expresses $\theta_t^*$ as a weighted average of Brownian increments over sampled trajectories, so the attacker needs only a simulator. For mitigation (Problem 2), the minimax KL control problem is shown to be equivalent to a risk-sensitive control problem (Problem 4) and a two-player zero-sum stochastic differential game (Problem 5); under a matching condition between the noise covariance $h_th_t^\top$ and the control-effectiveness matrix $g_t R_t^{-1} g_t^\top$, Theorems 3 and 4 turn the HJB and HJI equations into linear backward Chapman-Kolmogorov PDEs, and express the value and saddle-point policies as Monte Carlo expectations over the uncontrolled dynamics. The paper thus claims that worst-case stealthy attacks and their countermeasures are computable in real time for nonlinear systems without explicit models or an explicit policy synthesis step.
Load-bearing premise
The mitigation results stand on the matching condition that the noise covariance and the controller's effectiveness-weighted cost matrix must be proportional; if a system's noise and control directions do not align in that way, the path-integral defense formula is not derived.
Editorial extensions
If this is right
- Worst-case stealthy attacks for a fixed, known controller policy can be synthesized in real time from simulations, so systems whose models exist only as digital twins or neural ODEs remain amenable to principled attack analysis.
- Because the value function obeys the Feynman-Kac representation, the same attack synthesizer works for high-dimensional nonlinear systems where grid-based PDE solvers are infeasible, subject only to Monte Carlo sampling accuracy.
- The mitigation policy obtained from risk-sensitive control and the saddle-point policy obtained from the zero-sum game coincide, unifying two existing solution concepts under one path-integral computation.
- The optimal attack bias is deterministic in continuous time, in contrast to the randomized optimal attacks found in discrete-time linear settings; only the underlying noise remains random.
- When the matching condition holds, the controller can compute a saddle-point policy online from uncontrolled simulations of the system, without solving a game PDE.
Reading between the lines
- Editorial inference: the matching condition is structural: if $h_th_t^\top$ and $g_t R_t^{-1} g_t^\top$ act in different subspaces, no scalar $\alpha$ can satisfy Assumption 3, so the mitigation formulas are not justified for those systems. A testable extension is to lift the condition by allowing a state-dependent $\alpha$ or by adding a second controller channel.
- Editorial inference: the attack-synthesis side (Theorem 2) needs no matching condition, suggesting an asymmetry: worst-case attack analysis is more broadly applicable than the proposed mitigation formula, so a practical recipe would pair Theorem 2 with a different defense for non-matching systems.
- Editorial inference: the paper's own Remark 3 concedes that KL divergence is only an asymptotic measure of stealthiness; for short horizons its higher-order error behavior may diverge from true detectability, so the synthesized attacks and defenses should be interpreted through the lens of that surrogate.
- Editorial inference: equation (22) weights trajectories by $\exp(\frac{1}{\lambda}\int c)$, so for large costs the estimate concentrates on rare paths; variance reduction such as cross-entropy or scheduled sampling is a natural next step before deployment.
Signed reviews
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. This paper develops path integral methods for synthesizing and mitigating stealthy attacks in continuous-time nonlinear stochastic systems. The attacker's problem is formulated as a KL-constrained optimal control problem (Problem 1) and shown, via Girsanov's theorem, to be equivalent to a quadratic-cost stochastic optimal control problem (Theorem 1). The solution is then expressed as a path integral and computed by Monte Carlo sampling (Theorem 2). The mitigation problem is formulated as a minimax KL control problem (Problem 2), connected to risk-sensitive control and a two-player zero-sum stochastic differential game (Problems 4 and 5), and solved by path integral formulas under an added matching assumption (Theorems 3 and 4). The approach is illustrated on unicycle navigation and cruise control examples, with reported crash probabilities under attack and mitigation.
Significance. If the technical gaps are repaired, the paper would extend stealthy attack analysis from the discrete-time linear-Gaussian setting to a continuous-time nonlinear setting, and would demonstrate a sampling-based, simulator-driven method that does not require explicit policy synthesis. The Girsanov reduction in Theorem 1 and the Feynman-Kac steps are, in structure, standard and plausible, and the connections among KL minimax control, risk-sensitive control, and H-infinity control are useful. The paper also gives explicit Monte Carlo formulas, which is a strength for reproducibility in principle. The significance is, however, limited by the restrictive matching conditions (Assumptions 2 and 3) and by the rank-deficiency problems described below, which currently prevent the validation examples from being reproduced from the stated equations.
major comments (5)
- [Theorem 2, Eq. (22)] The matrix H_t = h_t^T (h_t h_t^T)^{-1} is undefined whenever h_t h_t^T is singular, and this is exactly the situation in both validation examples. In Section V-A the unicycle has h_t with two zero rows, so h_t h_t^T = diag(0,0,sigma^2,nu^2) is singular, and Section V-B has the same rank-deficient structure. The paper states no full-rank assumption and introduces no pseudoinverse, so the attack synthesis formula (25) and the reported P_crash values are not computable from the equations as written. Since actuators typically affect only a subset of state directions, this is a load-bearing gap for the paper's central claim.
- [Theorems 3 and 4, Eqs. (33), (43)-(44)] The controller formulas invert g_t R_t^{-1} g_t^T - (1/lambda) h_t h_t^T, but under Assumption 2 or 3 this matrix is proportional to h_t h_t^T, which is singular in both examples because g and h have zero rows for the unactuated state components. The paper neither restricts the inverse to the actuated subspace nor introduces a pseudoinverse. Moreover, the verification of Assumption 3 in Section V-A is performed on a 2x2 block rather than on the full 4x4 matrices appearing in the theorem, so the stated matrix equations are not actually satisfied by the validation systems. The mitigation policy (43) is therefore not justified for the systems used in the numerical experiments.
- [Assumptions 2 and 3] The mitigation results apply only when the noise covariance and the control-effectiveness covariance are matched, i.e., h_t h_t^T = xi g_t R_t^{-1} g_t^T or the equivalent Assumption 3. This is a genuine structural restriction: for arbitrary f, g, h it need not hold, and the abstract and title claim applicability to nonlinear CPS generally, which is stronger than what Theorems 3 and 4 establish. The paper should either state the scope explicitly as systems satisfying the matching condition or provide a relaxation with an error analysis.
- [Theorems 3 and 4 proofs, Eqs. (38) and (49)] In the Cole-Hopf linearization, the potential term is displayed as ell_t Psi_t / lambda in both proofs, but the stated Feynman-Kac solutions (39) and (50) use the exponents -1/gamma and -1/alpha, which require the potential coefficient to be ell_t/gamma and ell_t/alpha respectively. If the displayed lambda is not a typographical error, the proofs of Theorems 3 and 4 are internally inconsistent; this needs to be corrected or clarified.
- [Problem 5 / Theorem 4] The paper derives saddle-point policies from the HJI equation, but it does not provide a verification theorem showing that the resulting pair (u*, theta*) is a saddle point of the zero-sum game, nor does it state conditions such as the Isaacs condition or convexity-concavity of the Hamiltonian under which the game has a value. This is needed to support the claim that (43)-(44) solve Problem 2 rather than merely satisfy first-order necessary conditions.
minor comments (4)
- [Section III, Eq. (25)] The notation is ambiguous: the left-hand side of (25) is written as an approximation to theta*_t, while (22) states theta*_t dt equals an expression involving h_t dw_t; the relationship between dt and the sqrt(Delta t) discretization should be made explicit.
- [Section V-A] The verification of Assumption 3 is presented as a 2x2 matrix equation, but h and g are 4x2 matrices in the dynamics; the authors should explain the reduction to the 2x2 block and why the inverse formulas in Theorems 3 and 4 are well-defined in that setting.
- [Appendix] The statement that both alpha and beta can be made arbitrarily small simultaneously as h tends to zero is asserted from Figure 7; a short analytic argument or a precise limit would make the claim rigorous.
- [General] No code or data are provided for the numerical experiments; releasing the simulator and sampling scripts would substantially improve reproducibility and would help readers verify the reported P_crash values once the rank issues are addressed.
Circularity Check
No significant circularity: the main derivations are self-contained applications of Girsanov, Feynman-Kac, and Cole-Hopf transformations, with structural assumptions stated explicitly rather than smuggled in.
full rationale
The paper's claimed derivation chain is not circular. Theorem 1 reduces Problem 1 to the quadratic-cost SOC Problem 3 by an exact Girsanov computation: the KL divergence D(P||Q) is rewritten as (1/2) E_P[∫||θ_t||^2 dt], so the objective in (12) is obtained by identity, not by definitional fiat. Theorem 2 then applies the standard Cole-Hopf transformation and Feynman-Kac lemma to convert the HJB equation into a linear backward Chapman-Kolmogorov PDE, whose Monte Carlo evaluation is justified by the strong law of large numbers. The mitigation results in Theorems 3 and 4 are explicit dynamic-programming derivations in which Assumptions 2 and 3 are stated as sufficient structural matching conditions that make the nonlinear PDE linear; these are restrictive assumptions, but they are not hidden reuse of the conclusions. The self-citations [11], [13], and [33] support side claims or prior applicability of path integral methods, but the load-bearing proofs are reproduced in the present paper, and the cited results are published and independent enough not to constitute circularity. The numerical validation uses the paper's own formulas with stated parameter choices; the possible singularity of (h_t h_t^T) in the examples is a correctness or well-posedness concern about the examples, not a circularity, because no fitted parameter is renamed as a prediction and the reported P_crash values are not used as inputs to the derivation.
Assumptions & free parameters
free parameters (4)
- lambda (stealth/performance trade-off weight) =
2 and 0.1 (unicycle); 3 and 1.5 (cruise)
- Control cost weight R_t =
I (unicycle); 4 (cruise)
- Noise levels sigma_t, nu_t =
0.1 (unicycle); sqrt(0.005) (cruise)
- Cost weights b_t and eta_t for target and unsafe penalties =
b_t=0.1, eta_t=0.1 (unicycle); b_t=0.02, eta_t=0.02 (cruise)
assumptions (8)
- standard math Girsanov theorem applies, giving D(P||Q) = (1/2) E_P of the integral of ||theta_t||^2 for attack signals in class V.
- standard math Feynman-Kac lemma gives existence and uniqueness of the linearized value function PDE.
- standard math Strong law of large numbers justifies replacing expectations by finite-sample averages in Eqs. (25), (52), and (53).
- domain assumption The controller can monitor the disturbance process v_t directly and can halt the system when an anomaly is found.
- domain assumption Cost function c_t is quadratic in u_t: c_t = ell_t(x_t) + (1/2) u_t^T R_t u_t.
- ad hoc to paper Noise and control covariances are matched: h_t h_t^T = xi g_t R_t^{-1} g_t^T for some 0 < xi < lambda.
- ad hoc to paper Noise and control covariances satisfy h_t h_t^T = alpha (g_t R_t^{-1} g_t^T - (1/lambda) h_t h_t^T) for some alpha > 0.
- domain assumption Attack signals are restricted to drift perturbations with unit diffusion (Definition 1); diffusion attacks are dismissed via mutual singularity.
Cite this review
Pith. "Pith review of Path Integral Methods for Synthesizing and Preventing Stealthy Attacks in Nonlinear Cyber-Physical Systems." pith.science (2026). https://pith.science/paper/NQ7VJVH6
@misc{pith2026250417118,
author = {Pith},
title = {Pith review of: Path Integral Methods for Synthesizing and Preventing Stealthy Attacks in Nonlinear Cyber-Physical Systems},
year = {2026},
howpublished = {\url{https://pith.science/paper/NQ7VJVH6}},
note = {Machine review of arXiv:2504.17118}
}
read the original abstract
This paper studies the synthesis and mitigation of stealthy attacks in nonlinear cyber-physical systems (CPS). To quantify stealthiness, we employ the Kullback-Leibler (KL) divergence, a measure rooted in hypothesis testing and detection theory, which captures the trade-off between an attacker's desire to remain stealthy and her goal of degrading system performance. First, we synthesize the worst-case stealthy attack in nonlinear CPS using the path integral approach. Second, we consider how a controller can mitigate the impact of such stealthy attacks by formulating a minimax KL control problem, yielding a zero-sum game between the attacker and the controller. Again, we leverage a path integral-based solution that computes saddle-point policies for both players through Monte Carlo simulations. We validate our approach using unicycle navigation and cruise control problems, demonstrating how an attacker can covertly drive the system into unsafe regions, and how the controller can adapt her policy to combat the worst-case attacks.
Reference graph
Works this paper leans on
-
[1]
Cyber–physical systems: A perspective at the centennial,
K.-D. Kim and P. R. Kumar, “Cyber–physical systems: A perspective at the centennial,” Proceedings of the IEEE , vol. 100, no. Special Centennial Issue, pp. 1287–1308, 2012
2012
-
[2]
Special issue on cyber-physical systems [scanning the issue],
R. Poovendran, K. Sampigethaya, S. K. S. Gupta, I. Lee, K. V . Prasad, D. Corman, and J. L. Paunicka, “Special issue on cyber-physical systems [scanning the issue],” Proceedings of the IEEE, vol. 100, no. 1, pp. 6–12, 2011
work page 2011
-
[3]
Stuxnet worm impact on industrial cyber-physical system security,
S. Karnouskos, “Stuxnet worm impact on industrial cyber-physical system security,” in IECON 2011-37th Annual Conference of the IEEE Industrial Electronics Society . IEEE, 2011, pp. 4490–4494
work page 2011
-
[4]
Lessons learned from the maroochy water breach,
J. Slay and M. Miller, “Lessons learned from the maroochy water breach,” in International conference on critical infrastructure protection. Springer, 2007, pp. 73–82
work page 2007
-
[5]
Attack models and scenarios for networked control systems,
A. Teixeira, D. P ´erez, H. Sandberg, and K. H. Johansson, “Attack models and scenarios for networked control systems,” in Proceedings of the 1st international conference on High Confidence Networked Systems , 2012, pp. 55–64
work page 2012
-
[6]
Data-injection attacks in stochastic control systems: Detectability and performance tradeoffs,
C.-Z. Bai, F. Pasqualetti, and V . Gupta, “Data-injection attacks in stochastic control systems: Detectability and performance tradeoffs,” Automatica, vol. 82, pp. 251–260, 2017
work page 2017
-
[7]
The performance and limitations of ϵ-stealthy attacks on higher order systems,
E. Kung, S. Dey, and L. Shi, “The performance and limitations of ϵ-stealthy attacks on higher order systems,” IEEE Transactions on Automatic Control, vol. 62, no. 2, pp. 941–947, 2016
work page 2016
-
[8]
Worst-case stealthy innovation-based linear attack on remote state estimation,
Z. Guo, D. Shi, K. H. Johansson, and L. Shi, “Worst-case stealthy innovation-based linear attack on remote state estimation,” Automatica, vol. 89, pp. 117–124, 2018
work page 2018
Show all 35 references
-
[9]
Worst-case stealthy innovation-based linear attacks on remote state estimation under Kullback–Leibler diver- gence,
J. Shang, H. Yu, and T. Chen, “Worst-case stealthy innovation-based linear attacks on remote state estimation under Kullback–Leibler diver- gence,” IEEE Transactions on Automatic Control , 2021
2021
-
[10]
Secure networked control systems,
H. Sandberg, V . Gupta, and K. H. Johansson, “Secure networked control systems,” Annual Review of Control, Robotics, and Autonomous Systems, vol. 5, pp. 445–464, 2022. 13
2022
-
[11]
Simulator-driven deceptive control via path integral approach,
A. Patil, M. Karabag, T. Tanaka, and U. Topcu, “Simulator-driven deceptive control via path integral approach,” Submitted to 2023 IEEE Conference on Decision and Control (CDC) , 2023
2023
-
[12]
Stochastic game approach for replay attack detection,
F. Miao, M. Pajic, and G. J. Pappas, “Stochastic game approach for replay attack detection,” in 52nd IEEE conference on decision and control. IEEE, 2013, pp. 1854–1859
2013
-
[13]
Covert ve- hicle misguidance and its detection: A hypothesis testing game over continuous-time dynamics,
T. Tanaka, K. Sawada, Y . Watanabe, and M. Iwamoto, “Covert ve- hicle misguidance and its detection: A hypothesis testing game over continuous-time dynamics,” IEEE Control Systems Letters , 2024
2024
-
[14]
Path integrals and symmetry breaking for optimal control theory,
H. J. Kappen, “Path integrals and symmetry breaking for optimal control theory,” Journal of Statistical Mechanics: Theory and Experiment , vol. 2005, no. 11, p. P11011, 2005
2005
-
[15]
A generalized path integral control approach to reinforcement learning,
E. Theodorou, J. Buchli, and S. Schaal, “A generalized path integral control approach to reinforcement learning,” The Journal of Machine Learning Research, vol. 11, pp. 3137–3181, 2010
2010
-
[16]
Aggressive driving with model predictive path integral control,
G. Williams, P. Drews, B. Goldfain, J. M. Rehg, and E. A. Theodorou, “Aggressive driving with model predictive path integral control,” in2016 IEEE International Conference on Robotics and Automation (ICRA) . IEEE, 2016, pp. 1433–1440
2016
-
[17]
Model predictive path integral control: From theory to parallel computation,
G. Williams, A. Aldrich, and E. A. Theodorou, “Model predictive path integral control: From theory to parallel computation,” Journal of Guidance, Control, and Dynamics , vol. 40, no. 2, pp. 344–357, 2017
2017
-
[18]
Oksendal, Stochastic differential equations: an introduction with applications
B. Oksendal, Stochastic differential equations: an introduction with applications. Springer Science & Business Media, 2013
2013
-
[19]
Generalized Neyman-Pearson lemma via convex duality,
J. Cvitani ´c and I. Karatzas, “Generalized Neyman-Pearson lemma via convex duality,” Bernoulli, pp. 79–97, 2001
2001
-
[20]
Lattimore and C
T. Lattimore and C. Szepesv ´ari, Bandit algorithms . Cambridge Uni- versity Press, 2020
2020
-
[21]
T. M. Cover, Elements of information theory . John Wiley & Sons, 1999
1999
-
[22]
Lecture notes on information theory,
Y . Polyanskiy and Y . Wu, “Lecture notes on information theory,”Lecture Notes for ECE563 (UIUC) , 2014
2014
-
[23]
The optimal finite-sample error probabil- ity in asymmetric binary hypothesis testing,
V . Lungu and I. Kontoyiannis, “The optimal finite-sample error probabil- ity in asymmetric binary hypothesis testing,” in 2024 IEEE International Symposium on Information Theory (ISIT) . IEEE, 2024
2024
-
[24]
Relative entropy and free energy dualities: Connections to path integral and KL control,
E. A. Theodorou and E. Todorov, “Relative entropy and free energy dualities: Connections to path integral and KL control,” The 51st IEEE Conference on Decision and Control (CDC) , pp. 1466–1473, 2012
2012
-
[25]
Øksendal, Stochastic differential equations
B. Øksendal, Stochastic differential equations . Springer, 2003
2003
-
[26]
W. H. Fleming and H. M. Soner, Controlled Markov processes and viscosity solutions. Springer Science & Business Media, 2006, vol. 25
2006
-
[27]
R. F. Stengel, Optimal control and estimation . Courier Corporation, 1994
1994
-
[28]
Durrett, Probability: theory and examples
R. Durrett, Probability: theory and examples . Cambridge university press, 2019, vol. 49
2019
-
[29]
Risk-sensitive linear/quadratic/Gaussian control,
P. Whittle, “Risk-sensitive linear/quadratic/Gaussian control,” Advances in Applied Probability , vol. 13, no. 4, pp. 764–777, 1981
1981
-
[30]
Optimal stochastic linear systems with exponential per- formance criteria and their relation to deterministic differential games,
D. Jacobson, “Optimal stochastic linear systems with exponential per- formance criteria and their relation to deterministic differential games,” IEEE Transactions on Automatic control , vol. 18, no. 2, pp. 124–131, 1973
1973
-
[31]
Entropic value-at-risk: A new coherent risk measure,
A. Ahmadi-Javid, “Entropic value-at-risk: A new coherent risk measure,” Journal of Optimization Theory and Applications , vol. 155, pp. 1105– 1123, 2012
2012
-
[32]
Risk sensitive path integral control,
B. v. d. Broek, W. Wiegerinck, and H. Kappen, “Risk sensitive path integral control,” arXiv preprint arXiv:1203.3523 , 2012
2012 arXiv
-
[33]
Risk-minimizing two-player zero-sum stochastic differential game via path integral con- trol,
A. Patil, Y . Zhou, D. Fridovich-Keil, and T. Tanaka, “Risk-minimizing two-player zero-sum stochastic differential game via path integral con- trol,” in 2023 62nd IEEE Conference on Decision and Control (CDC) . IEEE, 2023, pp. 3095–3101
2023
-
[34]
An iterative method for nonlinear stochastic optimal control based on path integrals,
S. Satoh, H. J. Kappen, and M. Saeki, “An iterative method for nonlinear stochastic optimal control based on path integrals,” IEEE Transactions on Automatic Control , vol. 62, no. 1, pp. 262–276, 2016
2016
-
[35]
Bas ¸ar and P
T. Bas ¸ar and P. Bernhard,H-infinity optimal control and related minimax design problems: a dynamic game approach . Springer Science & Business Media, 2008
2008
Reviewed August 16, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.